Shortcuts on this page
r m x toggle line displays
j k next/prev highlighted chunk
0 (zero) top of page
1 (one) first highlighted chunk
1""" Routines for manipulating RFC2047 encoded words.
3This is currently a package-private API, but will be considered for promotion
4to a public API if there is demand.
6"""
7from __future__ import unicode_literals
8from __future__ import division
9from __future__ import absolute_import
10from future.builtins import bytes
11from future.builtins import chr
12from future.builtins import int
13from future.builtins import str
15# An ecoded word looks like this:
16#
17# =?charset[*lang]?cte?encoded_string?=
18#
19# for more information about charset see the charset module. Here it is one
20# of the preferred MIME charset names (hopefully; you never know when parsing).
21# cte (Content Transfer Encoding) is either 'q' or 'b' (ignoring case). In
22# theory other letters could be used for other encodings, but in practice this
23# (almost?) never happens. There could be a public API for adding entries
24# to the CTE tables, but YAGNI for now. 'q' is Quoted Printable, 'b' is
25# Base64. The meaning of encoded_string should be obvious. 'lang' is optional
26# as indicated by the brackets (they are not part of the syntax) but is almost
27# never encountered in practice.
28#
29# The general interface for a CTE decoder is that it takes the encoded_string
30# as its argument, and returns a tuple (cte_decoded_string, defects). The
31# cte_decoded_string is the original binary that was encoded using the
32# specified cte. 'defects' is a list of MessageDefect instances indicating any
33# problems encountered during conversion. 'charset' and 'lang' are the
34# corresponding strings extracted from the EW, case preserved.
35#
36# The general interface for a CTE encoder is that it takes a binary sequence
37# as input and returns the cte_encoded_string, which is an ascii-only string.
38#
39# Each decoder must also supply a length function that takes the binary
40# sequence as its argument and returns the length of the resulting encoded
41# string.
42#
43# The main API functions for the module are decode, which calls the decoder
44# referenced by the cte specifier, and encode, which adds the appropriate
45# RFC 2047 "chrome" to the encoded string, and can optionally automatically
46# select the shortest possible encoding. See their docstrings below for
47# details.
49import re
50import base64
51import binascii
52import functools
53from string import ascii_letters, digits
54from future.backports.email import errors
56__all__ = ['decode_q',
57 'encode_q',
58 'decode_b',
59 'encode_b',
60 'len_q',
61 'len_b',
62 'decode',
63 'encode',
64 ]
66#
67# Quoted Printable
68#
70# regex based decoder.
71_q_byte_subber = functools.partial(re.compile(br'=([a-fA-F0-9]{2})').sub,
72 lambda m: bytes([int(m.group(1), 16)]))
74def decode_q(encoded):
75 encoded = bytes(encoded.replace(b'_', b' '))
76 return _q_byte_subber(encoded), []
79# dict mapping bytes to their encoded form
80class _QByteMap(dict):
82 safe = bytes(b'-!*+/' + ascii_letters.encode('ascii') + digits.encode('ascii'))
84 def __missing__(self, key):
85 if key in self.safe:
86 self[key] = chr(key)
87 else:
88 self[key] = "={:02X}".format(key)
89 return self[key]
91_q_byte_map = _QByteMap()
93# In headers spaces are mapped to '_'.
94_q_byte_map[ord(' ')] = '_'
96def encode_q(bstring):
97 return str(''.join(_q_byte_map[x] for x in bytes(bstring)))
99def len_q(bstring):
100 return sum(len(_q_byte_map[x]) for x in bytes(bstring))
103#
104# Base64
105#
107def decode_b(encoded):
108 defects = []
109 pad_err = len(encoded) % 4
110 if pad_err:
111 defects.append(errors.InvalidBase64PaddingDefect())
112 padded_encoded = encoded + b'==='[:4-pad_err]
113 else:
114 padded_encoded = encoded
115 try:
116 # The validate kwarg to b64decode is not supported in Py2.x
117 if not re.match(b'^[A-Za-z0-9+/]*={0,2}$', padded_encoded):
118 raise binascii.Error('Non-base64 digit found')
119 return base64.b64decode(padded_encoded), defects
120 except binascii.Error:
121 # Since we had correct padding, this must an invalid char error.
122 defects = [errors.InvalidBase64CharactersDefect()]
123 # The non-alphabet characters are ignored as far as padding
124 # goes, but we don't know how many there are. So we'll just
125 # try various padding lengths until something works.
126 for i in 0, 1, 2, 3:
127 try:
128 return base64.b64decode(encoded+b'='*i), defects
129 except (binascii.Error, TypeError): # Py2 raises a TypeError
130 if i==0:
131 defects.append(errors.InvalidBase64PaddingDefect())
132 else:
133 # This should never happen.
134 raise AssertionError("unexpected binascii.Error")
136def encode_b(bstring):
137 return base64.b64encode(bstring).decode('ascii')
139def len_b(bstring):
140 groups_of_3, leftover = divmod(len(bstring), 3)
141 # 4 bytes out for each 3 bytes (or nonzero fraction thereof) in.
142 return groups_of_3 * 4 + (4 if leftover else 0)
145_cte_decoders = {
146 'q': decode_q,
147 'b': decode_b,
148 }
150def decode(ew):
151 """Decode encoded word and return (string, charset, lang, defects) tuple.
153 An RFC 2047/2243 encoded word has the form:
155 =?charset*lang?cte?encoded_string?=
157 where '*lang' may be omitted but the other parts may not be.
159 This function expects exactly such a string (that is, it does not check the
160 syntax and may raise errors if the string is not well formed), and returns
161 the encoded_string decoded first from its Content Transfer Encoding and
162 then from the resulting bytes into unicode using the specified charset. If
163 the cte-decoded string does not successfully decode using the specified
164 character set, a defect is added to the defects list and the unknown octets
165 are replaced by the unicode 'unknown' character \uFDFF.
167 The specified charset and language are returned. The default for language,
168 which is rarely if ever encountered, is the empty string.
170 """
171 _, charset, cte, cte_string, _ = str(ew).split('?')
172 charset, _, lang = charset.partition('*')
173 cte = cte.lower()
174 # Recover the original bytes and do CTE decoding.
175 bstring = cte_string.encode('ascii', 'surrogateescape')
176 bstring, defects = _cte_decoders[cte](bstring)
177 # Turn the CTE decoded bytes into unicode.
178 try:
179 string = bstring.decode(charset)
180 except UnicodeError:
181 defects.append(errors.UndecodableBytesDefect("Encoded word "
182 "contains bytes not decodable using {} charset".format(charset)))
183 string = bstring.decode(charset, 'surrogateescape')
184 except LookupError:
185 string = bstring.decode('ascii', 'surrogateescape')
186 if charset.lower() != 'unknown-8bit':
187 defects.append(errors.CharsetError("Unknown charset {} "
188 "in encoded word; decoded as unknown bytes".format(charset)))
189 return string, charset, lang, defects
192_cte_encoders = {
193 'q': encode_q,
194 'b': encode_b,
195 }
197_cte_encode_length = {
198 'q': len_q,
199 'b': len_b,
200 }
202def encode(string, charset='utf-8', encoding=None, lang=''):
203 """Encode string using the CTE encoding that produces the shorter result.
205 Produces an RFC 2047/2243 encoded word of the form:
207 =?charset*lang?cte?encoded_string?=
209 where '*lang' is omitted unless the 'lang' parameter is given a value.
210 Optional argument charset (defaults to utf-8) specifies the charset to use
211 to encode the string to binary before CTE encoding it. Optional argument
212 'encoding' is the cte specifier for the encoding that should be used ('q'
213 or 'b'); if it is None (the default) the encoding which produces the
214 shortest encoded sequence is used, except that 'q' is preferred if it is up
215 to five characters longer. Optional argument 'lang' (default '') gives the
216 RFC 2243 language string to specify in the encoded word.
218 """
219 string = str(string)
220 if charset == 'unknown-8bit':
221 bstring = string.encode('ascii', 'surrogateescape')
222 else:
223 bstring = string.encode(charset)
224 if encoding is None:
225 qlen = _cte_encode_length['q'](bstring)
226 blen = _cte_encode_length['b'](bstring)
227 # Bias toward q. 5 is arbitrary.
228 encoding = 'q' if qlen - blen < 5 else 'b'
229 encoded = _cte_encoders[encoding](bstring)
230 if lang:
231 lang = '*' + lang
232 return "=?{0}{1}?{2}?{3}?=".format(charset, lang, encoding, encoded)